Tweaking for perfect jerk/accel/pressure advance blob removal.
-
So, my issue started when I switched from a direct extruder (no gearing) to the E3D titan (3:1 gearing). At the end of each move, most notably on corners, a small blob would be made. Today I tried to eradicate this once and for all! (not an entire failure….). In the past, I've used various amounts of coast/wipe/retract settings to tweak this, as the blobs usually precede a retraction. This doesn't work as not all are made at a retraction though, some are actually made when the nozzle does a travel move (eg, from perimeter to infill). Coasting/wiping/extra distance on retract also has the downsides of messing with the layer alignment, as some parts of the layer would be partly underextruded (coast) or partly overextruded where I have upped extra distance after retraction to compensate. You can mess with these settings for YEARS and never get the perfect balance, there is always a compromise.... In the end, I'd rather find an area to hide retraction blobs, and keep good layer alignment. Add to this that the blobs only really show on 300 micron layers, I assume due to the pressure increases associated, then coasting / retract settings become tedious as you have to have different settings for each layer height as 1mm of coast will leave a lot of gaps on 100 micron layers.
So, with the idea that it was mostly due to slow travel moves, which may also be the reason for blobs on retract, I messed around with the settings....
Some background! My settings are usually....
I mostly print at 40 mm/s, 100mm/s travel moves. Cartesian machine, Bed moves in Y (weighs around 1.2kg).
M566 X600 Y400 Z8 E300 ; Jerk (mm/min)
M203 X10000 Y6000 Z300 E300 ; maximum speeds (mm/min)
M201 X3000 Y3000 Z100 E250 ; accelerations (mm/s^2)First off, I upped XY jerk, then E jerk, then Z jerk one at a time, to high amounts with no retractions on the outer perimeter. No difference at all!
Next up was a similar experiment with accel, again nothing. Next I set it back to originals and went for the pressure advance config, This worked well actually and brought down blobbing alot, but had its own drawbacks. In fact, it gave similar results to coast/wipe/retract tweaking where the section just after pressure advance took place would be slightly underextruded. However, it is nowhere near as bad and can be set similarly for all layer heights (I assume, I've not tried it yet...). So perhaps this is the compromise I can live with to remove blobbing??My next tests will be to have a lot more retractions, so that moving from perimeter to infill is not just a travel move but also a retract and then see if various jerk/acceleration settings affect the presence of blobs. So far, more E jerk has not.... What does everyone else do? Pressure advance seems like the best road to go down, But I can't help thinking its a work around to an underlying issue.
-
Just a couple of observations.
Comparing your extruder settings with mine (I also have E3D Titans - 3 of them in fact, feeding a diamond hot end) I'd say that maybe they are a bit low which would likely lead to slow retraction and might be part of the problem. You are using 300mm /min which is only 5mm /sec whereas I'm using 2400 mm/min (40mm/sec). Also, my accel is set to 1000 compare to your 250. So it might be worth upping you extruders speeds and accels to make retraction a bit snappier.
I also find that, contrary to what a lot of people say, I get better results with faster speeds - especially non print moves. You do need a very rigid machine though and with your bed moving in Y, you might be a bit restricted in how fast you can go. It might be worth upping speeds a bit, especially non-print moves just to see if it helps.
Also, have you tuned the retraction amount? I use the minimum retraction that I can print a couple of objects about 100mm apart without any stringing between them.
Oh, and temperature too - again I print a tower varying the temperature every 10mm or so, then pick the lowest temperature that still gives a good finish. This is almost invariably at the bottom end (or with T glass even lower) than is recommended but it works for me.
Just some ideas…..
HTH -
If you get small blobs on the print just before travel moves, this means you need more retraction. But I agree with deckingman, you also need to increase your extruder acceleration and maximum speeds a lot.
If you use a hot end that is sensitive to over-retraction such as the all metal E3D, then as you increase pressure advance you should reduce retraction.
-
Interesting Ian, I'm using for extruder (also titan/20mm nema 17 pancake stepper/24v)
Max speed 3600
Accel. 1000
Jerk 1200
Pressure advance 0.1 (with 17cm bowden tube)
Slic3r using (for ABS):
1.8mm retract (for e3d v6 genuine hotend)
50mm/s
And its working extremely well, certainly retracts very sharply with a definite "click" back and to, doesn't stall so its performing the moves, I do get a tiny bit of under extrude on infill just after a retract but I suspect I could dial it back to 1.6mm or lower and still get no blobs. -
DJ,
Very similar settings on extruders then. Haven't really played around much with pressure advance. Tried it once but had an awful lot of reverse extrudes going on even at 0.1. Wonder if it's anything to do with me using 3 extruders and maybe I should try 0.03? But then, maybe using firmware retraction (retracting all 3 filaments) interferes with pressure advance?
-
I'm running my direct drive 0.9 deg extruders at 9000 acceleration , 70 mm/s max speed with 25 mm/s jerk settings. I was running 90 mm/s jerk settings in the past and just got tired of the thap thap thap sound I was getting. I found 25 mm/s jerk setting gave me good results. I also use -0.05 mm on the preload or whatever it is to keep seam lines cleaner looking.
The titan however being geared is going to be much different. I would speed it way up and see where it starts to skip steps and then back it down a bit. Faster retractions appear to always be better short of loosing steps. I guess others will correct me if I'm wrong.
Jeff
-
If it can help I run my titan mounted with 1.8°degree motor on 24v at 9000 in acceleration and 2000 in jerk. I put this value because there are the max value trying on my machine, so no relation…? and it works fine !
-
@cubexupgrade:
If it can help I run my titan mounted with 1.8°degree motor on 24v at 9000 in acceleration and 2000 in jerk. I put this value because there are the max value trying on my machine, so no relation…? and it works fine !
More or less this makes the motor behave as if acceleration is turned off. It should snap very quickly on retracts. I also run 24v, can't see running a printer off 12v if 24v is an option. Setting the max speed of course is required or it'll chew the gears on the extruder really quickly. Keep in mind having the jerk setting that high might reduce the life of the extruder a bit. Some random thoughts, chillin out on a Friday night with an IPA.
Jeff
-
Might try mine a bit faster then.
Switched it to 3000 acceleration and 1500 (mm/min) jerk. Sounds click-ier but doesn't seem to alter the print quality that I can see. There must in theory be a point at which faster retraction does nothing to improve normal speed printing, but I am sure at extreme speeds it will help a lot.
-
You're right, the delrin gear will dead quickly, but e3d sell spare part and it's really fun when your printer doesn't stop moving (i can't see a pause for retract) !
If you're chilling time is not full you can watch a test done for a filament manufacturer : https://youtu.be/ZbAPCmVaDfA -
Just a couple of observations.
Comparing your extruder settings with mine (I also have E3D Titans - 3 of them in fact, feeding a diamond hot end) I'd say that maybe they are a bit low which would likely lead to slow retraction and might be part of the problem. You are using 300mm /min which is only 5mm /sec whereas I'm using 2400 mm/min (40mm/sec). Also, my accel is set to 1000 compare to your 250. So it might be worth upping you extruders speeds and accels to make retraction a bit snappier.
I also find that, contrary to what a lot of people say, I get better results with faster speeds - especially non print moves. You do need a very rigid machine though and with your bed moving in Y, you might be a bit restricted in how fast you can go. It might be worth upping speeds a bit, especially non-print moves just to see if it helps.
Also, have you tuned the retraction amount? I use the minimum retraction that I can print a couple of objects about 100mm apart without any stringing between them.
Oh, and temperature too - again I print a tower varying the temperature every 10mm or so, then pick the lowest temperature that still gives a good finish. This is almost invariably at the bottom end (or with T glass even lower) than is recommended but it works for me.
Just some ideas…..
HTHIt's not completely a retraction issue, a blob of similar size is created with both travel moves and retraction. As far as retraction goes though, my slicer is set to retract 1.8mm @ 55m/s in the slicer. I feel like I need to make the all the settings that bit snappier and the blobs will disappear without the need for pressure advance. As for temperature, It's set to maximize layer adhesion in my set up, which is usually 195-205c for PLA. This creates a bit of stringing, but I'd rather have the stringing and good layer adhesion. I've not tried temp calibration to see if blobbing is removed, but the blobbing appears on just about any material @ 300 microns. I print in PET, PLA and ABS. I find E3D's lack of support when it comes to settings on the Titan annoying. They've wrote a lot about settings for their hotends.
If you get small blobs on the print just before travel moves, this means you need more retraction. But I agree with deckingman, you also need to increase your extruder acceleration and maximum speeds a lot.
If you use a hot end that is sensitive to over-retraction such as the all metal E3D, then as you increase pressure advance you should reduce retraction.
I've found that pressure advance is not completely satisfactory as it creates areas of under and over extrusion. This is only very slight but enough to see with your own eyes on black PLA. In comparison, without pressure advance I get perfect layer adhesion with no noticeable undulations except for on travel moves and retractions with the blob. I will try with some of the settings posted here though and possibly post some pictures of my calibration cubes.
What I find interesting is that similar settings on my direct drive non geared extruder did not have these blobs on either retraction or travel moves, what changed when I switched to 3:1 gearing?
-
….......................What I find interesting is that similar settings on my direct drive non geared extruder did not have these blobs on either retraction or travel moves, what changed when I switched to 3:1 gearing?
Well you kind of answered your own question - the 3:1 gearing is what changed.
-
So another update,
I upped the extruder to -
Max speed 3600
Accel. 1000
Jerk 1200Some difference to the blobs, but only marginally. I still don't get how the 3:1 gearing can create such a difference, even when I up the settings. Do other people get such blobs? They start about 1.5mm away from a retraction point, and protrude roughly 0.1mm to make a highly visible blob.
The left is the old settings with pressure advance (M572 D0 S0.1), the middle is just the old settings and the right is the new settings. You can see that pressure advance has definitely stopped the blobs, but added under extrusion to a new line. The blobs on new and old settings look exactly the same to my eye, but the camera actually highlighted that they have a difference in size. -
I upped the XY acceleration and jerk too, along with the new extruder settings to see if this had an affect. It certainly made the printer snappier, but it had little affect on retraction blobbing.
XY accel - 5000 (upped from 3000)
XY Jerk - 1000, 1000 (upped from 600, 400)
-
Ah now those pictures paint a different story altogether. Assuming the side is facing upwards, then it doesn't happen every layer. I'd guess form the pictures every 4th or 5th layers. To me, that would indicate something wrong mechanically.
Edit. Which no amount of tweaking of accelerations, jerk or speed is going to cure. I'd look for something loose or sticking somewhere that might have happened when you changed the extruder. -
It is happening every 3 layers, as it retracts on a different corner each time. This is 'optimize start points' feature in S3D. I can confirm that every blob is a retraction point in the slicer simulation. The edge without retractions is completely smooth. The reason its set this way is so that every blob can be seen, rather then having a merged edge of blobs. Normal prints I will choose a hidden place for these. I've uploaded a picture of an edge with no retractions (left side) so you can see what good layer alignment should look like without a blob caused by travel movements or retractions.
-
Sorry I'm getting a bit confused because you said a few posts back that it's not completely a retraction issue. However you now say that every blob is a retraction point which would indicate that it is. The other thing is that you also said that at the temperature you like to print, you do get some stringing which is really why we use retraction. You don't have to put up with stringing just to get good layer adhesion. So, it seems to me that maybe you should increase the retraction amount to prevent the stringing, which will probably cure the blobs.
-
Sorry, I'll try to be more clear.
These cubes all have blobs created by retraction. I know this as I've set retraction to take place when it moves from perimeter to infill. However, If I turn this retraction off a blob is still created by the travel move. Initially I thought it was solely a retraction issue.
What confuses me is, why does having retraction have such little effect on the blobs? If this was solely a retraction issue, then surely even bad retraction settings should have an effect.
Temperature may have some affect on blob size, but I doubt it will remove the issue entirely. I will however do some tests just so it is another removed variable.
-
Do you print external perimeters first or the other way around. This is a very odd problem, I wonder if you print internal perimeters first whether you will still get these artefacts but you wont care as you wont see them?
-
Increase the E jerk setting. I had been getting that since I boosted the acceleration for the x & y, and never did change the jerk settings. I originally had it at a really low value of 20, and changing it to 200 practically got rid of it. I'm using Titan bowden setup to a 1.75mm e3vd6 with a pressure advance of 0.03.
